sktime/sktime

[ENH] interfaces to mobility/trajectory related estimators - scikit-mobility, movingpandas

Open

#4259 aperta il 22 feb 2023

Vedi su GitHub
 (2 commenti) (0 reazioni) (0 assegnatari)Python (1192 fork)batch import
enhancementgood first issuemodule:detectionmodule:transformations

Metriche repository

Star
 (7162 star)
Metriche merge PR
 (Merge medio 26g 10h) (86 PR mergiate in 30 g)

Descrizione

From discord discussion with Yuri Skylakhter.

Mobility or trajectory analysis has a rich collection of feature extractors that are specific to 2D (plane or spherical) and 3D trajectories. The estimators here are mostly transformations and annotation estimators.

This may be interesting for users of sktime, e.g., combining feature extraction and time series classification, etc.

Two packages that collect transformations, annotation, clustering etc estimators:

https://github.com/scikit-mobility/scikit-mobility https://github.com/movingpandas/movingpandas

(looke like nice packages!)

The first general question would be: how to deal with this? We could interface the estimators, perhaps best starting with transformations, as sktime estimators, with an optional dependency on the above packages.

We could also talk to the maintainers of the above packages and see whether interface homogenization is sth they might like to consider.

Guida contributor